Human Visual System and Wavelet Transform Based Disparity Estimation

Abstract

In this paper, three new simple and efficient disparity estimation algorithms are proposed which are based on wavelet transform and human visual system. The core idea behind the proposed algorithms is to create a set of different left and right images followed by the disparity estimation from the different set. The obtained disparity from different left and right images are merged into a single more accurate disparity using human visual system model. Experiments with test and real stereo image pairs show the validity of the proposed method.